Ryan Wang

I think what you'll see: - A lot of great companies with AI agents - and the line blurred between inbound support and "customer experience" (help page + home page) - A new crop of copilot tools - harder, higher value tickets are going to people and there's a lot of "work around the work" (e.g. collecting data across internal tools, wrap up notes) - Economics coming into play - many AI agents are still more expensive than BPO agents in places like Colombia & Philippines - A resurgence of "classic" contact center tools - workforce management, voice of the customer, quality assurance all retooled to contemplate AI agents Support was already a massive software market before AI automation, and there will be many huge companies

@HarryStebbings @HarryStebbings You missed @assembledhq! Of all the AI platforms, we're the only ones with the orchestration layer to route & plan across human + AI agents. Support teams large (Robinhood, Stripe) and small (EliseAI, Flexcar) have automation *and* work that still goes to people.

One market I cannot figure out in a world of AI is customer support. You have: Sierra Decagon GigaML Parloa Forethought Kore Cresta All of these have raised over $100M. and then oldies like Intercom Zendesk Help me out, how does this market shake out?

just used @NotionHQ AI (instead of search) for the first time. works as advertised, which i consider to be magical. i wonder what it means for the RAG wrappers that don't own the generative interface.

@michelleefang from @assembledhq thursday may 8, a panel with ex-cto of notion, head of api eng at openai, & stripe's longest-tenured engineer on ai adoption in engineering lu.ma/b80npkpn?tk=H1…

if you made it to here: 1) retweet or like to help others in SF 🔁 2) drop any other events I may have missed in the comments - thank you for making the SF tech community so vibrant! 🙏

Plugging you into 40+ SF tech events (Config + Stripe Sessions edition!) ⚡️ A list of what's happening this week (May 5 - May 12) ⬇️

refreshing my read-every-issue list on ai. currently: - interconnects.ai by @natolambert - importai.substack.com by @jackclarkSF - oneusefulthing.org by @emollick any recs?

The first rule of daylight savings. There is always another daylight savings edge case you haven't accounted for yet.

The full podcast is a gem covering AI pricing, defensibility in a copilot world, and evolving business models. I feel lucky to work with Jake on a daily basis, and still learned something new. Worth a listen: youtube.com/watch?v=vdbMuq…

Well, the secret's out. @jakesaper just revealed on 20vc how we made him participate in a mock board meeting at 11pm, right after his honeymoon, as part of @assembledhq's Series A process. But seven years in, selecting board partners was one of the most consequential decisions we've ever made.

Awesome to see @assembledhq recognized on the @Forbes 2025 Best Startup Employers List We ask people to build the company they want to work at, and I'm super proud of the collective effort 🏆 forbes.com/lists/americas…

Oldie but goodie, @AcquiredFM episode on TSMC. Still amazed that Morris Chang started the company at 55, among other nuggets. acquired.fm/episodes/tsmc-…

First thought re: @SlackHQ outage: "We'll be so productive without distractions!" Reality: I'm texting people to get stuff done, inefficient and not searchable (and text is even worse at threads...) Truth is, in a company, people need to communicate to get stuff done. Tools just facilitate this need.

Every few weeks, when a new model drops, I find myself re-sharing this classic: x.com/gdb/status/173… Especially pertinent at the application layer. If you’ve done your evals well, easy gains. If you haven’t, you’ll probably be chasing your tail in circles.

Been traveling a lot lately. Got back to San Francisco and realized wow, I am living in the future. It's easy to forget when you're in it every day.

RIP Bram Moolenar I felt a lot of imposter syndrome in my first software engineering job. Learning to use vim made me feel like a “real” coder. The only true text editor!

Oppenheimer sequel but it's just about his brother founding the Exploratorium

@VictorPontis D’oh, missed it the first time. Maybe biased! I think I saw it in the biography “The Pope of Physics.” I’ll need to dig this up…

@wgyn_ Source here — twitter.com/VictorPontis/s… Lol is this Fermi talk your Chicago bias coming thru?

Ernest Rutherford trained at least **11** Nobel prize winning scientists. This is also kind of a cool graphic since Rayleigh, Thomson, and Rutherford all have a form of physical scattering named after them. So this chart shows the scattering / chain reaction of their teaching!

Saw this product strategy outline on a desk @assembledhq 🧠 In all seriousness, we’re hiring! (Including for a team focused on our AI-powered support bot product). DMs are open if interested.

I haven't had Prime in years. With free shipping, I still tend to (not always, but pretty often) get things in 2-days. Curious why. I assume it's operationally simpler to do 2-day for the entire supply chain?

@maccaw Thanks for sharing this, Alex. And congrats! 🙇‍♂️

Getting my US green card was one of the hardest things I've accomplished. It was especially difficult because I didn't go to university. I would like to share my story on how I achieved this, in the hope that it may be useful to other immigrants. Obviously a big disclaimer here - INAL. You *will* need a good lawyer to help you navigate all of this. Once you have decided you want to move to the US, you'll need a job and a visa. Typically the job part of this is the easier part - companies are desperate for talent. Getting a visa, on the other hand, can be extremely hard. Not just because of the paperwork - but because it's a mental mind game (more on that later). I first went for a O1 visa. This is a temporary visa that you need to renew every year. It requires evidence of 'extraordinary ability' - in practice if you are a talented engineer with a good track record it is possible to get. The most important part (for me) was that there wasn't a hard requirement on university qualifications. I'm a high school dropout and, while I don't regret that choice, that makes it much harder to get a visa. To get an O1 you'll need a job offer, a good lawyer, and a bunch of evidence (such as high profile open-source work, published materials, talks at conferences, high salary, and letters of recommendation). To bolster my published work I decided to write a programming book (which ended up being published by O'Reilly). It took roughly a year to get all this together. And then a few months to get an answer. Typically startups won't bother sponsoring H1Bs (because the annual cap / lottery on them basically makes them untenable), but the smart ones will sponsor O1s, especially if the founder is an immigrant and familiar with the process. It'll cost them ~$5-10k in legal fees (not much for an engineer). You will need the founder involved to help connect you to notable people (who can hopefully recommend you). You will need to do a lot of the prep work yourself - don't expect your lawyer to do it. You will need to understand the ins and outs of the system. I was fortunate - I got an offer from Twitter and they had an excellent legal team. The previous year of work writing the book paid off. My O1 was approved and I could start working in the US. Pro tip - make sure you wait 10 days or so after entering the US to apply for your SSN. It takes a while for all their systems to talk to each other. The biggest downside of an O1 is it's connected to an employer. I.e. you can't go and start a company (which is silly). Other downsides are that if you are married, your partner is not allowed to work. And of course an O1 isn't a path to permanent residency by itself - it's a temporary visa. It's tough to describe the mental torment of this process to those unfamiliar. Every year, I would visit the US embassy in London, a daunting gloomy place, to renew my visa. Looking back, this process was straightforward, but in the moment, it was terrifying. I was being interviewed by an agent who could, on a whim, deny my visa, thereby uprooting my entire life. I'd be back at square one - my life reset. You would hear horror stories of people who had their lives turned upside down because of a minor error in their paperwork or a misunderstanding during the interview. It felt like walking a tightrope. So, after 4 years in the states I decided to go for something a bit more permanent: the EB1 green card. This is basically the O1 "extraordinary alien" equivalent - but much much harder. You have to demonstrate a lot of evidence (highly notable in your field, or started notable a company, raised lots of $$$, etc etc). It's often denied (as my first application was). My stack of evidence was longer than any book I'd written. You have to attend an nerve-racking interview. But eventually my lawyer and I made our case, and we were notified my green-card was on its way. That was one of the best days of my life! So that's my story. Many moons ago I wrote an article on visas, which I believe is still accurate - if you're a talented immigrant looking to move to the states, that's a good next read. underdog.io/blog/engineers… In conclusion, I wish all of this were easier. At the risk of tooting my own horn, it's fair to say that individuals like myself are a net positive for the U.S. I've already paid more taxes than the average American will in their lifetime, created hundreds of American jobs, and contributed to a country in which I can't even vote. Of course, I'm incredibly grateful for the opportunities the U.S. has provided me. And, if I were starting out today, I believe I would still try to move to America. However, with each passing year, countries like Canada and the U.K., with their modern visa programs, seem increasingly appealing to the next generation of ambitious entrepreneurs.